Agents in Economic Markets and Games
169
assess their strategies based on the fitness in the prisoner’s dilemma game
(Table 6.5).
The strategy played in the prisoner’s dilemma game was a 16-state strategy. A 16-state strategy had a similar structure to the design of the automaton
discussed by Miller [132].
Table 6.8 represents a three-state automaton represented as a series of
strings showing a three state strategy. Figure 6.25 displays the corresponding
strategy of this automaton. The starting state is State 0. In this state the
player will cooperate. If the other player cooperates the player will move to
the State 1, else it will move to State 2. Depending on the new states its next
moves will depend on what is represented in the state it is currently in.
TABLE 6.8: Example of a three state machine represented by automaton.
State C/D
Next State, if Other
Player Cooperates
Next State, if Other
Player Defects
0
C
1
2
1
D
0
2
2
C
2
2
C
( S t a t e 0 )
C
( S t a t e 2 )
D
( S t a t e 1 )
c
c
d
c , d
d
FIGURE 6.25: Example of automaton represented by Table 6.8.
The automaton used in the FLAME iterated prisoner’s dilemma game
uses a 16-state strategy. A 16-state strategy is represented using 4 bits for
each state. Each strategy will contain 16 states; a payoff playing that strategy
is the score. Players maintain a database of these strategies in their memory, to
aid their competition in the simulation. The structure of the strategy database
is pictured in Figure 6.26.
Figure 6.27 shows the structure of one state in this strategy. The state in
the strategy is a string of 9 bits. The first bit represents which strategy to
play when in this state. In Figure 6.27, the player will cooperate in this state.
After doing so, depending on what the other player plays, it will move to a
new state. If the other player cooperates, the player will move to the next
169
assess their strategies based on the fitness in the prisoner’s dilemma game
(Table 6.5).
The strategy played in the prisoner’s dilemma game was a 16-state strategy. A 16-state strategy had a similar structure to the design of the automaton
discussed by Miller [132].
Table 6.8 represents a three-state automaton represented as a series of
strings showing a three state strategy. Figure 6.25 displays the corresponding
strategy of this automaton. The starting state is State 0. In this state the
player will cooperate. If the other player cooperates the player will move to
the State 1, else it will move to State 2. Depending on the new states its next
moves will depend on what is represented in the state it is currently in.
TABLE 6.8: Example of a three state machine represented by automaton.
State C/D
Next State, if Other
Player Cooperates
Next State, if Other
Player Defects
0
C
1
2
1
D
0
2
2
C
2
2
C
( S t a t e 0 )
C
( S t a t e 2 )
D
( S t a t e 1 )
c
c
d
c , d
d
FIGURE 6.25: Example of automaton represented by Table 6.8.
The automaton used in the FLAME iterated prisoner’s dilemma game
uses a 16-state strategy. A 16-state strategy is represented using 4 bits for
each state. Each strategy will contain 16 states; a payoff playing that strategy
is the score. Players maintain a database of these strategies in their memory, to
aid their competition in the simulation. The structure of the strategy database
is pictured in Figure 6.26.
Figure 6.27 shows the structure of one state in this strategy. The state in
the strategy is a string of 9 bits. The first bit represents which strategy to
play when in this state. In Figure 6.27, the player will cooperate in this state.
After doing so, depending on what the other player plays, it will move to a
new state. If the other player cooperates, the player will move to the next